An apparatus and corresponding method are provided for encoding video signal data by applying an adaptive quantizer parameter responsive to bit-rate; where a video encoder for encoding video signal data for an image block relative to at least one reference picture includes an adaptive quantizer for applying an adaptive quantizer parameter responsive to a current bit-rate; and where a corresponding method includes initializing an adaptive quantizer, calculating the bits used by the video signal data with the adaptive quantizer, assigning bits to each of a plurality of picture types, calculating a difference in adaptive quantizer responsive to the assigned bits from the previous number of bits used by the group of pictures and obtaining an updated adaptive quantizer for the video signal data, calculating an adaptive quantizer for a portion of the video signal data, encoding the portion of the video signal data, and obtaining the actual number of bits used for the encoded portion of the video signal data.
Legal claims defining the scope of protection. Each claim is shown in both the original legal language and a plain English translation.
1. An apparatus, comprising: a video encoder for encoding video signal data for a group of pictures (GOP) formed of I, B and P pictures, the encoder comprising an input for receiving the video signal data and an adaptive quantizer including a hardware processor for applying an adaptive quantizer parameter responsive to a current bit-rate, wherein the hardware processor of the adaptive quantizer is configured to: calculate a bit budget for encoding at least one item from within the GOP, wherein the calculation (i) determines a number of bits for the GOP, (ii) allocates bits for I, P, and B pictures in the GOP, (iii) determines a quantizer parameter for the I picture by dividing a quantizer parameter of a previous GOP by a constant representative of the number of respective P and B pictures included in the GOP and adding the result to a differential quantizer parameter, (iv) determines a quantizer parameter for a P picture by adding a quantizer parameter for a previous P picture to a differential quantizer parameter for a P picture, and (v) calculates a quantizer parameter for a B picture, by adding a quantizer parameter for a previous B picture to a differential quantizer parameter for a B picture; and wherein the quantizer parameter is responsive to said calculated bit budget.
A video encoder apparatus encodes video data by adaptively adjusting the quantization parameter based on the current bitrate for a group of pictures (GOP) containing I, B, and P frames. The encoder calculates a bit budget for encoding the GOP. This involves determining the total number of bits for the GOP and allocating bits to I, P, and B frames. It determines the quantizer parameter for the I frame by dividing the quantizer parameter of the previous GOP by a constant related to the number of P and B frames in the current GOP, and then adding a differential quantizer parameter. It determines the quantizer parameter for a P frame by adding the quantizer parameter for the previous P frame to a P-frame differential quantizer parameter. Similarly, it calculates the B-frame quantizer parameter. The quantizer parameter is dynamically adjusted in response to the calculated bit budget.
2. The apparatus of claim 1 , wherein the length of the image sequence is unknown.
The video encoder apparatus of claim 1, which encodes video data by adaptively adjusting the quantization parameter based on the current bitrate for a group of pictures (GOP) containing I, B, and P frames, where the encoder calculates a bit budget for encoding the GOP, is further specified where the length of the video sequence to be encoded is unknown. This means the encoder can handle video streams of indefinite duration without pre-defined length limitations.
3. The apparatus of claim 1 , wherein the group of pictures comprises pictures between two adjacent intra coded pictures and including the first intra coded picture.
The video encoder apparatus of claim 1, which encodes video data by adaptively adjusting the quantization parameter based on the current bitrate for a group of pictures (GOP) containing I, B, and P frames, where the encoder calculates a bit budget for encoding the GOP, is further specified where the GOP includes all frames between two adjacent intra-coded (I) frames, inclusive of the first I frame. This defines the GOP structure as containing the pictures residing between two I-frames, useful for scene changes and random access points.
4. The apparatus of claim 1 , wherein an initial bit budget ratio for I, P, and B types of pictures is I:P:B=6:3:1.
The video encoder apparatus of claim 1, which encodes video data by adaptively adjusting the quantization parameter based on the current bitrate for a group of pictures (GOP) containing I, B, and P frames, where the encoder calculates a bit budget for encoding the GOP, is further specified where the initial bit allocation ratio for I, P, and B frames within the GOP is set to I:P:B = 6:3:1. This represents a weighting for the relative importance of each frame type, with I-frames receiving the largest share of bits.
5. The apparatus of claim 1 , wherein an initial bit budget ratio for the group of pictures is initialized to a bit budget ratio resulting from a previous group of pictures.
The video encoder apparatus of claim 1, which encodes video data by adaptively adjusting the quantization parameter based on the current bitrate for a group of pictures (GOP) containing I, B, and P frames, where the encoder calculates a bit budget for encoding the GOP, is further specified where the initial bit budget ratio for the current GOP is initialized based on the bit budget ratio resulting from the previous GOP. This enables temporal adaptivity in bit allocation between GOPs by using prior encoding statistics to inform the initial bit budget distribution.
6. A method for encoding a group of pictures at a video encoder including an input for receiving the group of pictures and providing the group of pictures to an adaptive quantizer, comprising, at a hardware processor of the adaptive quantizer: calculating a bit budget for encoding at least one item from within said group of pictures (GOP), wherein the calculating comprises: determining the number of bits for said GOP; allocating corresponding number of bits to I, P and B pictures in the GOP; adjusting a quantizer parameter responsive to said calculated bit budget; determining a quantizer parameter for said I picture by dividing a quantizer parameter of a previous group of pictures by a constant representative of the number of respective P and B pictures in the GOP and adding the result to a differential quantizer parameter; determining a quantizer parameter for a P picture, by adding a quantizer parameter for a previous P picture to a differential quantizer parameter for a P picture; determining a quantizer parameter for a B picture, by adding a quantizer parameter for a previous B picture to a differential quantizer parameter for a B picture; and iteratively repeating said calculating and adjusting to further adjust said quantizer parameter.
A video encoding method encodes a group of pictures (GOP) by adaptively quantizing the video data. It calculates a bit budget for encoding the GOP by determining the total number of bits for the GOP and allocating bits to I, P, and B frames. The quantizer parameter is adjusted based on the calculated bit budget. The quantizer parameter for the I frame is determined by dividing the quantizer parameter of the previous GOP by a constant representing the number of P and B frames and adding a differential quantizer parameter. The quantizer parameter for the P frame is determined by adding the quantizer parameter for the previous P frame to a P-frame differential quantizer parameter. Similarly, the B-frame quantizer parameter is determined. This calculation and adjustment process is repeated iteratively to fine-tune the quantizer parameter.
7. A method as defined in claim 6 , the group of pictures comprising pictures between two adjacent intra coded pictures and including the first intra coded picture.
The video encoding method as defined in claim 6, which encodes a group of pictures (GOP) by adaptively quantizing the video data, is further specified where the GOP comprises pictures between two adjacent intra coded pictures and including the first intra coded picture. This defines the GOP structure as containing the pictures residing between two I-frames, useful for scene changes and random access points.
8. A method as defined in claim 6 wherein the length of the group of pictures is unknown.
The video encoding method as defined in claim 6, which encodes a group of pictures (GOP) by adaptively quantizing the video data, is further specified where the length of the video to be encoded is unknown. The method can handle streams without predefined lengths.
9. The method of claim 6 , wherein an initial bit budget ratio for I, P, and B types of pictures is I:P:B=6:3:1.
The video encoding method of claim 6, which encodes a group of pictures (GOP) by adaptively quantizing the video data, is further specified where an initial bit budget ratio for I, P, and B types of pictures is I:P:B=6:3:1. This represents a weighting for the relative importance of each frame type, with I-frames receiving the largest share of bits.
10. The method of claim 6 , wherein an initial bit budget ratio for the group of pictures is initialized to a bit budget ratio resulting from a previous group of pictures.
The video encoding method of claim 6, which encodes a group of pictures (GOP) by adaptively quantizing the video data, is further specified where an initial bit budget ratio for the group of pictures is initialized to a bit budget ratio resulting from a previous group of pictures. This enables temporal adaptivity in bit allocation between GOPs by using prior encoding statistics to inform the initial bit budget distribution.
11. An apparatus for encoding a group of pictures comprising an adaptive quantizer including a hardware processor, said hardware processor being configured to: calculate a bit budget for encoding at least one item from within said group of pictures (GOP), wherein calculating the bit budget comprises: determining the number of bits for said GOP; allocating corresponding number of bits to I, P and B pictures in the GOP; adjust a quantizer parameter responsive to said calculated bit budget by determining a quantizer parameter for said I picture by dividing a quantizer parameter of a previous group of pictures by a constant representative of the number of respective P and B pictures in the GOP and adding the result to a differential quantizer parameter; determining a quantizer parameter for a P picture, by adding a quantizer parameter for a previous P picture to a differential quantizer parameter for a P picture; determining a quantizer parameter for a B picture, by adding a quantizer parameter for a previous B picture to a differential quantizer parameter for a B picture; and iteratively repeat said calculating and adjusting to further adjust said quantizer parameter.
An apparatus for video encoding adaptively quantizes video data for a group of pictures (GOP). The apparatus calculates a bit budget for the GOP, determining the total number of bits for the GOP and allocating bits to I, P, and B frames. A quantizer parameter is adjusted based on the calculated bit budget. The quantizer parameter for the I frame is determined by dividing the quantizer parameter of a previous GOP by a constant related to the number of P and B frames in the current GOP and then adding a differential quantizer parameter. The P-frame quantizer parameter is determined by adding the quantizer parameter for the previous P frame to a P-frame differential quantizer parameter. Similarly, the B-frame quantizer parameter is determined. The calculation and adjustment process repeats iteratively to fine-tune the quantizer parameter.
12. The apparatus of claim 11 , wherein the group of pictures comprises pictures between two adjacent intra coded pictures and including the first intra coded picture.
The video encoding apparatus of claim 11, which adaptively quantizes video data for a group of pictures (GOP), is further specified where the GOP comprises pictures between two adjacent intra coded pictures and including the first intra coded picture. This defines the GOP structure as containing the pictures residing between two I-frames, useful for scene changes and random access points.
13. The apparatus of claim 11 , wherein an initial bit budget ratio for I, P, and B types of pictures is I:P:B=6:3:1.
The video encoding apparatus of claim 11, which adaptively quantizes video data for a group of pictures (GOP), is further specified where an initial bit budget ratio for I, P, and B types of pictures is I:P:B=6:3:1. This represents a weighting for the relative importance of each frame type, with I-frames receiving the largest share of bits.
14. The apparatus of claim 11 , wherein an initial bit budget ratio for the group of pictures is initialized to a bit budget ratio resulting from a previous group of pictures.
The video encoding apparatus of claim 11, which adaptively quantizes video data for a group of pictures (GOP), is further specified where an initial bit budget ratio for the group of pictures is initialized to a bit budget ratio resulting from a previous group of pictures. This enables temporal adaptivity in bit allocation between GOPs by using prior encoding statistics to inform the initial bit budget distribution.
15. A non-transitory processor readable storage medium readable by machine, storing a computer program product comprising program code instructions executable by the machine for encoding a group of pictures, said program code instructions: calculating a bit budget for encoding at least one item from within said group of pictures (GOP), wherein the calculating comprises: determining the number of bits for said GOP; allocating corresponding number of bits to I, P and B pictures in the GOP; and adjusting a quantizer parameter responsive to said calculated bit budget by: determining a quantizer parameter for said I picture by dividing a quantizer parameter of a previous group of pictures by a constant representative of the number of respective P and B pictures in the GOP and adding the result to a differential quantizer parameter; determining a quantizer parameter for a P picture, by adding a quantizer parameter for a previous P picture to a differential quantizer parameter for a P picture; determining a quantizer parameter for a B picture, by adding a quantizer parameter for a previous B picture to a differential quantizer parameter for a B picture; and iteratively repeating said calculating and adjusting to further adjust said quantizer parameter.
A non-transitory computer-readable storage medium stores instructions to perform adaptive quantization for video encoding. The instructions calculate a bit budget for a group of pictures (GOP), determining the total number of bits and allocating bits to I, P, and B frames. The quantizer parameter is adjusted based on the calculated bit budget. The quantizer parameter for the I frame is determined by dividing the quantizer parameter of a previous GOP by a constant related to the number of P and B frames in the current GOP and adding a differential quantizer parameter. The P-frame quantizer parameter is determined by adding the quantizer parameter for the previous P frame to a P-frame differential quantizer parameter. Similarly, the B-frame quantizer parameter is determined. This calculation and adjustment process is repeated iteratively to refine the quantizer parameter.
16. The non-transitory processor readable storage medium of claim 15 , wherein the group of pictures comprises pictures between two adjacent intra coded pictures and including the first intra coded picture.
The computer readable storage medium of claim 15, which stores instructions to perform adaptive quantization for video encoding, is further specified where the group of pictures comprises pictures between two adjacent intra coded pictures and including the first intra coded picture. This defines the GOP structure as containing the pictures residing between two I-frames, useful for scene changes and random access points.
17. The non-transitory processor readable storage medium of claim 15 , wherein an initial bit budget ratio for I, P, and B types of pictures is I:P:B=6:3:1.
The computer readable storage medium of claim 15, which stores instructions to perform adaptive quantization for video encoding, is further specified where an initial bit budget ratio for I, P, and B types of pictures is I:P:B=6:3:1. This represents a weighting for the relative importance of each frame type, with I-frames receiving the largest share of bits.
18. The non-transitory processor readable storage medium of claim 15 , wherein an initial bit budget ratio for the group of pictures is initialized to a bit budget ratio resulting from a previous group of pictures.
The computer readable storage medium of claim 15, which stores instructions to perform adaptive quantization for video encoding, is further specified where an initial bit budget ratio for the group of pictures is initialized to a bit budget ratio resulting from a previous group of pictures. This enables temporal adaptivity in bit allocation between GOPs by using prior encoding statistics to inform the initial bit budget distribution.
Cooperative Patent Classification codes for this invention. Click any code to explore related patents in that topic.
June 15, 2005
June 20, 2017
Browse 5M+ US patents with plain-English claim translations and AI-generated analysis.